Robots to the Rescue
ASTM Subcommittee, Competitions and New NIST Facility Support Robot Standards
ASTM International Subcommittee E54.08 on Operational Equipment Collaborates to Produce Robot Performance Standards
Success came with a robot named Quince. In the days following March 11, 2011, when a major earthquake rocked northern Japan, caused a massive tsunami and led to the Fukushima nuclear disaster, responders came, and so did robots from nearly a dozen different organizations. A ground robot and a ground-aerial team assisted with earthquake damage inspection, and five marine vehicles were used for tsunami recovery operations — including the location and removal of downed boats leaking fuel into the surrounding fishing waters. Two unmanned aerial vehicles and at least six different ground robots were deployed at the Fukushima nuclear power plant.
Other robots had not been able to map the altered plant nor detect radiation plumes above the ground floor. Enter Quince. Though refined through tests and knowledge gained at robotics competitions, it was still considered to be a research prototype. Quince was the product of work led by Satoshi Tadokoro, Dr.Eng., a professor at Tohoku University, Japan, and president of the International Rescue System Institute. Quince’s independent front and rear flippers and wide body tracks enable advanced mobility across complex terrains, including the ability to climb steep steps with occlusions. When fitted with sensors to detect radiation plumes and a tether for communications, Quince successfully negotiated and mapped the upper stories of the disabled reactor building, providing essential information to emergency managers.
Standing behind Quince were standards from ASTM International E54.08.01 on Operational Equipment; Robots, part of Committee E54 on Homeland Security Applications.
E54.08.01 chairman Adam Jacoff, test director at the Intelligent Systems Division of the National Institute of Standards and Technology, Gaithersburg, Md., co-founded the RoboCup-Rescue League and its competitions in 2000 with Tadokoro, also an E54 member. Jacoff says that if there were no RoboCup-Rescue competitions and no ASTM International standards, there would have been no Quince. (For more about RoboCup-Rescue, see “Competition and Training Lead to Innovation,” below.)
The Response Robot
A response robot, also called an emergency response robot, provides support and protection during crisis situations to first responders and other on-site personnel.
Operated remotely and intended to perform specific tasks, a robot serves as an extension of the responder to explore and work in unsafe areas.
Robot research, development and construction is an international industry. In the United States, civilian and military groups are hard at work advancing the technology, as are universities and manufacturers in countries such as Canada, Germany, Israel, Japan, Thailand and Turkey.
Response robots consist of advanced technologies and systems, but they share certain basic characteristics: sensors, endurance, radio communications. At a crisis site, a robot should:
- Deploy rapidly,
- Allow reliable remote operation from a safe distance,
- Have the ability to move quickly though complex environments,
- Have enough protection to avoid being damaged while performing its duties,
- Be cost-effective and field-serviceable, and
- Be available in varying sizes and adaptable to varying environments.
Performance and Process
ASTM’s standards work on robots is about performance and process.
In the early 2000s, Jacoff was working on common tests that could be used to address responders’ robot needs such as rubble and terrain that a robot needs to navigate, or how to build a throwable robot, one that will, for example, withstand the force needed to land it on a roof and show whether an area is occupied or not. Jacoff had also been running RoboCupRescue competitions, which brought together the robotics research and development community to test their robots’ capabilities.
In fiscal year 2004, the U.S. Department of Homeland Security initiated a NIST project to consider responder capabilities and their robotics requirements, intended to lead to comprehensive standards that would support robotic technologies. The initiative to develop standards that would evaluate various performance attributes was begun in Subcommittee E54.08 on Operational Equipment.
The subcommittee’s mandate is to develop standards for operational equipment used for chemical, biological, radiological, nuclear and explosive — or CBRNE — incident response.
“We look at equipment and materials that might slip by the other subcommittees and provide the structure for the development of new standards and related documents,” says Philip Mattson, E54.08 chairman and acting standards executive, DHS.
The standards from E54.08 address requirements for communications, human-robot interfaces, logistics, manipulation, mobility, sensors and safety for different types of robots. “We are not looking to develop specifications for a standard robot of any kind,” says Jacoff. “Rather, we are developing standards to test robots through a consensus process with robot developers, responders, and civilian and military test administrators.”
To date, E54.08 has guided 18 new standards from the drawing board through the approval process; 14 more are work items moving through the consensus process and even more will follow. Current standards address how to evaluate robotic capabilities — hurdles, stairs and ramps; maneuvering and mobility; and communications, both in line of sight and out of line of sight. Manufacturers and user groups can then cite the standards relevant to robots for a particular application.
Particularly significant for most robots are the radio communications standards, E2854 and E2855.1 The communications standards development illustrates a crucial part of all of E54.08’s work – use of knowledge coming from actual robot trials. To do a real-world test of the claimed uninterrupted radio capability, robots had to navigate behind a monolithic structure and go into a “radio shadow” while continuing to communicate with operators. That trial showed a dubious, if nonexistent, communications link. Engineers tackled the problem, agreed on an appropriate protocol to bounce the signal effectively, tested the approach repeatedly and developed the ASTM standards based on that work.
Jacoff says that robotic competitions continually inform the work the of the E54.08 robotics task group. “The competition gives back to the standards process,” he says. “They [the researchers] are the ones who feed the process the most.” Hundreds of researchers, professors or Ph.D. candidates, graduate students and undergrads make up the teams entering robots in competition. Regional finals culminate in an international championship with the best of the best demonstrating how their robots can perform the same task 10 times over, and that proving ground gives performance data that validates the ASTM standards. It’s also cost-effective, as one facility would not be able to build the robots and test arenas and perform the numerous repetitions necessary to validate a test.
The NIST Robot Test Facility
In April, the National Institute of Standards and Technology opened the doors of its new Robot Test Facility in Gaithersburg, Md. The 10,000 square foot (930 square meter) facility has been designed to provide a wide range of standard terrains and obstacles to test the abilities of response robots.
“Emergency responders and soldiers need quantitative ways to measure whether mobile robots are capable and reliable enough to perform envisioned application tasks,” says Jacoff. “Our job at the new test center will be to develop standard test methods that facilitate comparisons of different robot configurations based on statistically significant robot performance data.”
The E54.08 tests developed address robot requirements from DHS and military organizations. These test methods, collectively known as the DHS-NIST-ASTM International Standard Test Methods for Response Robots, will be used to quantitatively evaluate robotic capabilities such as mobility and maneuvering; manipulation; power consumption; communications abilities; logistics and human-robot interaction. The test methods address a range of robot sizes, configurations and capabilities including throwable robots for reconnaissance tasks, mobile manipulator robots for package size and vehicle-born improvised explosive devices, along with rapidly deployable aerial and aquatic systems.
Robot capability data captured with standard test methods help measure system improvements, highlight breakthrough capabilities, guide procurement and deployment decisions, and support operator proficiency training. As more standards are developed and approved, they will be added to the test package.
The NIST Robot Test Facility will also provide support for the Defense Advanced Research Projects Agency’s Robotics Challenge, an initiative designed to promote innovation and spur development in robotic technology for disaster-response operations.
“DARPA recognizes that our suite of DHS-NIST-ASTM International Standard Test Methods for Response Robots already addresses several of the elemental capabilities necessary to complete their program and wants to use them to quantitatively evaluate teams,” says Jacoff. “This work will provide an excellent opportunity to extend and disseminate our test methods. The project will also include interlaboratory experiments to establish the “reproducibility” of the standard test methods using robots that span the design spectrum.
Robots in Action
“Whenever a natural disaster occurs or a bomb goes off, the local authorities call for a response robot to help them deal with situations that are potentially unsafe for human workers,” says Mattson.
“In addition to the humanitarian use of the robots for response and recovery,” says Robin Murphy, Ph.D., director of the Center for Robot-Assisted Rescue at Texas A&M University, College Station, Texas, “we have learned many lessons for the scientific community that will help improve the performance, sensors, and interfaces for these robots and make robots easier to use by anyone from anywhere in the world over the Internet.”
“In addition to urban search and rescue missions, robots have proven to be very capable platforms for bomb retrieval and disposal,” says Mattson.
Last year, Mattson and Jacoff attended the National Bomb Squad Commanders conference in New Mexico to demonstrate robot test methods to more than 300 bomb squad commanders and special guests.
“It’s heartening that they totally understand our project, how our suite of standard test methods helps them directly articulate their needs to robot developers, measure incremental progress and highlight new robotic capabilities, support procurements and acceptance testing, and then measure operator proficiency during responder training,” Jacoff says.
“Bomb squads and military IED technicians approach known explosives every day,” says Jacoff. “When robots fail, these men get in padded suits and walk up to known explosives — that has to stop. As we make more advancements in our technology, we’re finally getting close to a point where that may be possible.”
For More Information
1. E2854, Test Method for Evaluating Emergency Response Robot Capabilities: Radio Communication: Line-of-Sight Range, and E2855, Test Method for Evaluating Emergency Response Robot Capabilities: Radio Communication: Non-Line-of-Sight Range.Kessel Nelson is a freelance writer whose work has appeared in national and international publications, and he has covered diverse subjects, including art, energy, crime, science and health issues. He has a B.A. in history from the University of Pennsylvania and spends his time between Philadelphia and New York City.
Cicely Enright is associate editor of ASTM Standardization News.
Competition and Training Lead to Innovation
The use of response robots in times of trouble is well documented. But before a robot becomes available commercially, its technologies may have been tested multiple times in the proving grounds of such competitions that are part of the RoboCup Rescue or training exercises at Disaster City.
The RoboCupRescue project was inspired by the Great Hanshi-Awaji earthquake that hit Kobe City, Japan, in January 1995, resulting in more than 6,500 casualties and damages exceeding $1 billion dollars.
The lessons learned from the Hanshin-Awaji earthquake concluded that information systems should be built with the following requirements:
- Collection, accumulation, relay, selection, summarization and distribution of necessary information.
- Prompt support for planning disaster mitigation, search and rescue.
- Reliability and robustness of the system during routine and emergency operations.
Given the above requirements, the RoboCupRescue project promotes research and development in this socially significant domain at various levels involving multi-agent teamwork coordination, physical robotic agents for search and rescue, information infrastructures, personal digital assistants, a standard simulator and decision support systems, evaluation benchmarks for rescue strategies and robotic systems that are all integrated into a comprehensive systems in future.
Regional competitions highlight the work of thousands of researchers vying to be part of the final international competition, which showcases the top five teams from each region.
Located in College Station, Texas, Disaster City is a 52-acre facility that trains emergency response professionals in the skills and techniques that they need. Created by the Texas A&M Engineering Extension Service, the mock community features full-scale, collapsible structures designed to simulate various levels of disaster and wreckage including rubble piles, transportation accidents and downed buildings.
Emergency responders from around the globe bring their robots to Disaster City for unparalleled search and rescue training and exercises. While responders come to Disaster City to train, they also gain opportunities to learn more about developing robot technologies, voice their needs and learn more about the role of standards in robot performance.
This article appears in the issue of Standardization News.